Richard, actually it's 7 mA per 6.8 kOhm resistor for a total of about 14 mA as the "short circuit" current. A preamp or supply should be able tolerate that for a few seconds or minutes on one channel at least; accidental short circuits do occasionally happen. However, that requires a more robust construction than is found in many pieces of portable equipment that I've tested.
For P48 microphones, the DIN EN standard gives 10 mA as the maximum. Above 7 mA, more energy is spent heating up the phantom resistors than actually reaches the microphone, and many lightweight preamps or phantom supplies fall out of regulation. So most manufacturers don't go into that zone of diminishing returns. Nonetheless there are the inevitable wacky exceptions: many Earthworks models draw the full 10 mA, and many CAD Equitek models draw 8 mA.
Most modern P48 condenser microphones draw less than 6 mA apiece. Shure KSM-series microphones are typically specified at 5.5 mA or thereabouts. Schoeps CMC 5-- and 6-- preamps normally draw about 4.5 mA apiece from a 48 Volt supply, but the CMC 6-- amplifier, if it senses too low a supply voltage, may switch over to its 12-Volt mode and try to draw 10 mA. (It isn't a "12 through 48 Volt" amplifier; it's a dual-mode amplifier which requires either standard 12 Volt phantom powering with adequate current or standard 48 Volt phantom powering with adequate current.) Neumann TLM and fet 100 mikes typically draw 2 - 3 mA.
Neumann's older fet 80 designs (KM 84, U 87, etc.) draw less than 1 mA, as did Schoeps' pre-Colette-series CMT 5-series microphones. But the age of a design isn't a reliable guide. To many people's surprise, the original AKG C 451 from the 1970s--not the electret "reissue" (cough, cough)--tries to draw 6 mA from a standard 48 Volt supply. It really was a P12 microphone with a tolerance for 48 Volts, not a 48-Volt microphone as such.
A hint: If you're revamping a cheap phantom power supply, check its 6.8 kOhm supply resistors; make sure that the pair on each input is matched as closely as possible, since any imbalance there reduces the preamp's common mode rejection. Also, the closer the match, the better your protection against any hum or hash from the DC supply getting into the audio.
The specification says that the two resistors in any one pair (i.e. on any one input) should not differ from one another by more than 0.4%. Even expensive "audiophile grade" 1% resistors may thus be out of spec by as much as a factor of 5--a real case of misplaced priorities, which is unfortunately fairly common. The absolute 6.8 kOhm value has a 20% tolerance on it; having equal value resistors within each pair is far more important.
--best regards